Skip to contents

GLM sediment-temperature parameters from observed temperature profiles

Source: R/calc_sed_temp.R
calc_sed_temp.Rd

Fits an annual sediment-temperature cycle (sed_temp_mean, sed_temp_amplitude, sed_temp_peak_doy) for each GLM sediment zone from observed water-column temperature profiles. Used internally by build_aeme() to populate the GLM &sediment block, and exported so the same per-zone values can be generated as a calibration parameter table.

Usage

calc_sed_temp(
  aeme = NULL,
  obs_temp = NULL,
  hypsograph = NULL,
  sed_zones = NULL,
  max_depth = NULL,
  temp_var = "HYD_temp",
  nml_file = "glm4.nml",
  output = c("parameters", "nml", "summary"),
  depth_grid = 0.25,
  depth_tol = 2,
  min_obs = 10,
  min_months = 6,
  default_mean = 12,
  default_amplitude = 4,
  default_peak_doy = 46,
  borrow_amp_factor = 0.6,
  borrow_lag_days = 25,
  verbose = TRUE
)

Arguments

aeme

Aeme object; when supplied, obs_temp and hypsograph are extracted from it unless passed explicitly. Default NULL.

obs_temp

long df: Date, var_aeme, depth, value (degC), depths in m below surface. Optional when aeme is given.

hypsograph

df with depth (m, +down) or elev, and area (m2), used to area-weight within a zone and (with sed_zones = NULL) to derive the zones. Optional when aeme is given. NULL -> uniform weighting and sed_zones / max_depth must be supplied.

sed_zones

numeric vector of zone upper-boundary heights above the bed, ascending - exactly as passed to GLM zone_heights. Zone 1 is 0 -> sed_zones[1] (the deepest slab); the last value may exceed max depth. NULL (default) -> estimate_sed_zones() on hypsograph.

max_depth

max lake depth (m). NULL -> from hypsograph, else deepest obs (warns).

temp_var

value of var_aeme to keep.

nml_file

character; name of the GLM nml file the parameters belong to, used for the file column of the parameter table. Default "glm4.nml".

output

one of "parameters" (default) - a long AEME model-parameter table in param_colnames()(incl_opt = FALSE) order, ready for calibration; "nml" - a named list of the three per-zone vectors to splice into glm_nml[["sediment"]]; or "summary" - the per-zone diagnostic data.frame with parameters / nml / zone_series attributes.

depth_grid

within-zone integration step (m).

depth_tol

a cast is used for a zone only if its deepest sample is within depth_tol m of the top of that zone's depth band.

min_obs, min_months

thresholds for a harmonic fit (else monthly-range).

default_mean, default_amplitude, default_peak_doy

last-resort values. default_peak_doy = 46 is mid-Feb (S. hemisphere summer); use ~209 for N.

borrow_amp_factor, borrow_lag_days

when an unfitted deep zone borrows a cycle from the nearest shallower fitted zone: damp amplitude by borrow_amp_factor^(zone steps) and lag peak_doy by borrow_lag_days per step.

verbose

logical; print the assembled per-zone table, nml snippet and parameter table.

Value

Depends on output:

  • "parameters" - data.frame with columns model, file, name, value, min, max, group, index; one row per zone for each of sediment/sed_temp_mean, sediment/sed_temp_amplitude and sediment/sed_temp_peak_doy.

  • "nml" - named list sed_temp_mean / sed_temp_amplitude / sed_temp_peak_doy, each a per-zone numeric vector in GLM zone order.

  • "summary" - the per-zone diagnostic data.frame (zone 1 = deepest), carrying the "parameters" table, a ready-to-paste "nml" snippet and the per-zone "zone_series" as attributes.

Details

Inputs can be supplied directly (obs_temp, hypsograph, sed_zones) or pulled from an Aeme object via aeme: observations come from get_obs() and the hypsograph from input(). sed_zones defaults to estimate_sed_zones() on the hypsograph when not given.